Harden launch papercuts and clean builds

README.md

@@ -249,7 +249,7 @@ If you want a completely fresh start: `make clean` then `make dev`.
 | Data Collection | [TinyFish](https://www.tinyfish.ai?utm_source=github&utm_medium=organic&utm_campaign=bigset-developer-2026q2) APIs (Search, Fetch, Browser) |
 | AI orchestration | [Mastra](https://mastra.ai) workflows + [Vercel AI SDK](https://sdk.vercel.ai) + [OpenRouter](https://openrouter.ai) → Claude Sonnet (schema inference + populate agent) |
 | Table view | [TanStack Table](https://tanstack.com/table) + [react-window](https://github.com/bvaughn/react-window) virtualization |
-| Exports | CSV (built-in) + XLSX ([SheetJS](https://sheetjs.com), dynamic-imported) |
+| Exports | CSV (built-in) + XLSX ([write-excel-file](https://www.npmjs.com/package/write-excel-file), dynamic-imported) |
 | Analytics | [PostHog](https://posthog.com) — events, session replay, error tracking (optional) |
 
 ## 📁 Project Structure

frontend/app/dashboard/page.tsx

@@ -26,18 +26,18 @@ export default function DashboardPage() {
   const mine = useQuery(
     api.datasets.listMine,
     isAuthenticated ? {} : "skip",
-  );
+  ) as DatasetCardData[] | undefined;
   const showCurated = !isLocalMode;
   const curated = useQuery(
     api.datasets.listPublic,
     showCurated ? {} : "skip",
-  );
+  ) as DatasetCardData[] | undefined;
 
   // Quota limits are cloud-only. Local mode can create datasets without this gate.
   const usage = useQuery(
     api.quota.getMy,
     !isLocalMode && isAuthenticated ? {} : "skip",
-  );
+  ) as { remaining: number } | undefined;
   const atLimit = !isLocalMode && usage !== undefined && usage.remaining === 0;
 
   // Fire dashboard_viewed once per mount when both queries have resolved,
@@ -61,7 +61,7 @@ export default function DashboardPage() {
 
   const { filteredMine, filteredCurated } = useMemo(() => {
     const q = search.trim().toLowerCase();
-    const apply = (list: typeof mine) =>
+    const apply = (list: DatasetCardData[] | undefined) =>
       (list ?? []).filter((ds) =>
         !q ||
         ds.name.toLowerCase().includes(q) ||
@@ -187,7 +187,7 @@ export default function DashboardPage() {
           eyebrow="Yours"
           heading="Datasets you own"
           isLoading={mine === undefined}
-          datasets={filteredMine as unknown as DatasetCardData[]}
+          datasets={filteredMine}
           emptyState={
             search
               ? `No datasets of yours match "${search}".`
@@ -205,7 +205,7 @@ export default function DashboardPage() {
               eyebrow="Curated by BigSet"
               heading="Explore live datasets"
               isLoading={curated === undefined}
-              datasets={filteredCurated as unknown as DatasetCardData[]}
+              datasets={filteredCurated}
               emptyState={
                 search
                   ? `No curated datasets match "${search}".`

frontend/app/dataset/[id]/page.tsx

@@ -6,7 +6,7 @@ import { useCallback, useEffect, useMemo, useRef, useState } from "react";
 import { useMutation, useQuery } from "convex/react";
 import { api } from "@/convex/_generated/api";
 import type { Id } from "@/convex/_generated/dataModel";
-import { DatasetTable } from "@/components/table";
+import { DatasetTable, type DatasetMeta, type DatasetRow } from "@/components/table";
 import { useSelection } from "@/components/table/use-selection";
 import { SideSheet, CellDetail } from "@/components/SideSheet";
 import type { DatasetColumn } from "@/components/table/types";
@@ -23,6 +23,15 @@ import {
 import type { ProfileUser } from "@/lib/profile-user";
 import { useAppAuth, useAppClerk, useAppConvexAuth, useAppUser } from "@/lib/app-auth";
 
+type DatasetDetail = DatasetMeta & {
+  _id: Id<"datasets">;
+  ownerId: string;
+  rowCount?: number;
+  maxRowCount?: number;
+  seedKey?: string;
+  visibility?: "public" | "private";
+};
+
 export default function DatasetPage() {
   const params = useParams();
   const { isLoading: authLoading, isAuthenticated } = useAppConvexAuth();
@@ -49,11 +58,11 @@ export default function DatasetPage() {
   const dataset = useQuery(
     api.datasets.get,
     authLoading ? "skip" : { id: datasetId },
-  );
+  ) as DatasetDetail | undefined;
   const rows = useQuery(
     api.datasetRows.listByDataset,
     authLoading ? "skip" : { datasetId },
-  );
+  ) as DatasetRow[] | undefined;
   const updateRefreshSettings = useMutation(api.datasets.updateRefreshSettings);
   const updateMaxRowCount = useMutation(api.datasets.updateMaxRowCount);
   const usage = useQuery(
@@ -168,7 +177,7 @@ export default function DatasetPage() {
   }
 
   async function handleUpdate() {
-    if (!dataset || updating || dataset.status === "building" || dataset.status === "updating") return;
+    if (!dataset || !rows || updating || dataset.status === "building" || dataset.status === "updating") return;
     // A new run is starting — discard any lingering stop-latch from the previous run.
     setStopping(false);
     setUpdating(true);

frontend/app/dataset/new/page.tsx

@@ -55,6 +55,23 @@ function mapBackendColumn(col: InferredColumn, index: number): ProposedColumn {
   };
 }
 
+function getSchemaValidationError(datasetName: string, columns: ProposedColumn[]): string | null {
+  if (!datasetName.trim()) return "Dataset name is required.";
+  if (columns.length === 0) return "Add at least one column.";
+
+  const seenColumnNames = new Set<string>();
+  for (const column of columns) {
+    const normalizedName = column.name.trim().toLowerCase();
+    if (!normalizedName) return "Every column needs a name.";
+    if (seenColumnNames.has(normalizedName)) {
+      return `Column names must be unique. "${column.name.trim()}" is duplicated.`;
+    }
+    seenColumnNames.add(normalizedName);
+  }
+
+  return null;
+}
+
 function TypeSelector({ value, onChange }: { value: ColumnType; onChange: (v: ColumnType) => void }) {
   return (
     <div className="relative inline-flex">
@@ -172,6 +189,12 @@ export default function NewDatasetPage() {
 
   async function handleConfirm() {
     if (isCreating) return;
+    const validationError = getSchemaValidationError(datasetName, columns);
+    if (validationError) {
+      setError(validationError);
+      return;
+    }
+
     const maxRowCount = Number(maxRowCountInput);
     if (!Number.isInteger(maxRowCount) || maxRowCount < 1) {
       setError("Max rows must be a whole number greater than 0.");
@@ -186,20 +209,22 @@ export default function NewDatasetPage() {
     setIsCreating(true);
     setError(null);
     let datasetId: string;
+    const normalizedColumns = columns.map((column) => ({
+      name: column.name.trim(),
+      type: column.type,
+      description: column.description.trim() || undefined,
+      isPrimaryKey: column.isPrimaryKey || undefined,
+    }));
+
     try {
       datasetId = await createDataset({
-        name: datasetName,
-        description: prompt,
+        name: datasetName.trim(),
+        description: prompt.trim(),
         refreshCadence,
         maxRowCount,
-        columns: columns.map((c) => ({
-          name: c.name,
-          type: c.type,
-          description: c.description || undefined,
-          isPrimaryKey: c.isPrimaryKey || undefined,
-        })),
+        columns: normalizedColumns,
         retrievalStrategy: retrievalStrategy ?? undefined,
-        sourceHint: sourceHint || undefined,
+        sourceHint: sourceHint.trim() || undefined,
       });
     } catch (err) {
       const message = err instanceof Error ? err.message : "Failed to create dataset";
@@ -214,7 +239,7 @@ export default function NewDatasetPage() {
     try {
       track(EVENTS.DATASET_CREATED, {
         datasetId,
-        column_count: columns.length,
+        column_count: normalizedColumns.length,
         refreshCadence,
         maxRowCount,
       });

frontend/app/page.tsx

@@ -14,7 +14,7 @@ const PUBLIC_GRID_COUNT = 9;
 
 export default function Home() {
   const { isAuthenticated, isLoading } = useAppConvexAuth();
-  const publicDatasets = useQuery(api.datasets.listPublic, {});
+  const publicDatasets = useQuery(api.datasets.listPublic, {}) as DatasetCardData[] | undefined;
 
   // Fire once when the landing page actually displays to an anonymous
   // visitor. Skip if we'll immediately redirect them to the dashboard.
@@ -103,7 +103,7 @@ export default function Home() {
                 {shownDatasets.map((ds) => (
                   <DatasetCard
                     key={ds._id}
-                    dataset={ds as unknown as DatasetCardData}
+                    dataset={ds}
                   />
                 ))}
               </div>